!exclusive! | Marvelocity Pdf

\subsection{Learning the Residual} Define the residual speed: \begin{equation} \Delta V = V_{\text{SOG}} - V_{\text{HM}}, \end{equation} where $V_{\text{SOG}}$ is the measured speed over ground from AIS. We train a Gradient‑Boosted Regression Tree (XGBoost \cite{Chen2016}) to predict $\Delta V$ from the feature vector $\mathbf{x}$: \[ \mathbf{x} = \bigl[\,\underbrace{L, B, D, C_B}_{\text{design}};\, \underbrace{V_{\text{HM}}}_{\text{baseline}};\, \underbrace{U_{10}, \theta_{\text{wind}}}_{\text{wind}};\, \underbrace{H_s, \theta_{\text{wave}}}_{\text{wave}};\, \underbrace{U_c, \theta_{\text{current}}}_{\text{current}}\,\bigr]. \]

\subsection{Hybrid Strategies} Hybrid schemes—e.g., residual learning on top of HM \cite{Zhang2023}—have shown promise but often require vessel‑specific fine‑tuning. MarVelocity differentiates itself by learning a **universal correction** that transfers across ship types. marvelocity pdf

The final **MarVelocity** prediction is: \begin{equation} V_{\text{MarV}} = V_{\text{HM}} + \hat{\Delta V}(\mathbf{x}). \end{equation} Compared with the baseline speed‑keeping policy, the fleet

\subsection{Fuel‑Efficiency Gains} A six‑month field trial (January–June 2025) on a fleet of 150 container ships employed MarVelocity to compute \emph{optimal speed profiles} under real‑time weather forecasts. Compared with the baseline speed‑keeping policy, the fleet realized an average fuel reduction of **4.8 \%** (≈ 1.9 million kg CO\textsubscript{2} avoided). Compared with the baseline speed‑keeping policy

\section{Results} \label{sec:results} \subsection{Prediction Accuracy} Table~\ref{tab:accuracy} summarizes error metrics on the held‑out test fleet (150 vessels, 1.1 M observations).

\subsection{Ablation Study} Figure~\ref{fig:ablation} shows the impact of removing each environmental group from the feature set. Wind contributes the most to error reduction (ΔMAE = 0.04 knot), followed by waves (0.03 knot) and currents (0.02 knot).